- From: Dave Kristol <dmk@bell-labs.com>
- Date: Sat, 1 Mar 1997 17:51:25 -0500
- To: Larry Masinter <masinter@parc.xerox.com>
- Cc: state@xent.w3.org, http-wg@cuckoo.hpl.hp.com
At 8:38 AM -0800 3/1/97, Larry Masinter wrote: >This morning, I started having some really bad second >thoughts about the state of cookies. Here are some >hard questions (intended to be provocative): > >a) if we're not going to be compatible, why call it a "cookie" > at all? It's not all that descriptive. "Set-state" and > "state", for example. Strangely enough, I was having similar thoughts. I agree with your argument in favor of a change. The argument in favor of Set-Cookie2[/Cookie2] is to convey similarity of concept and behavior. > >b) if we're not going to be compatible, what's the rush? > > There's no installed base of "Set-Cookie2", so > why should we try to revise this quickly? My reason is personal. I would like to have finished the specification. However, if you believe there is merit in having an IETF specification for cookies, then sooner is better. I think the arguments for such a specification are: - the new specification is tighter and more likely to yield interoperable implementations - the new specification emphasizes the importance of privacy protections > >c) no-one has expressed an opinion that the new draft is > actually technically better than RFC 2109, only that > we should do this because of compatibility with some > part of the "installed base". What is "the new draft"? There is no new draft filed with IETF. I announced to a sub-group mailing list a draft for their inspection. I had not opened up that specification to http-wg. (Yes, I posted a state-mgmt-06 draft for people to inspect, but it only included errata to RFC 2109. It included none of the more radical changes, like a change of header names. And I never sent it to IETF.) > > It's a bad precedent to revise a proposed standard > for non-technical reasons, and only because one of > the implementors didn't actually read the spec at > the right time. I'll let Yaron Goland argue that one.... > > >d) There was some argument initially that the whole idea > of cookies was a bad idea. If we reset to proposed, will > we need to revisit those arguments, since what we're doing > with cookies is no longer a reflection of "running code". Aren't we still *at* Proposed? What's to reset? The delay will be in seeing two or more interoperating implementations. Only the vendors can tell us how long it would take to produce the browsers. I am not particularly happy at the recent turn of events, but I'm interested in producing a spec. that will actually get honored, not ignored. So I am willing to change the words to produce something people will use. Dave Kristol
Received on Saturday, 1 March 1997 14:58:44 UTC